Data security is an important issue in the age of big data. The existing data security approaches should be improved to cover inactive databases, i.e. the databases with existing information only, and suit the requirements of big data mining. Therefore, this paper proposes framework to protect the data anonymity in big data environment. The framework is mainly implemented in three steps: mining the association rules, computing the confidence of each rule, and determining the sensitivity of each rule using fuzzy logic. To process massive data, the authors paid attention to enhance the parallelism and scalability of the proposed framework. The proposed framework was verified through experiments on two datasets. Judging by metrics like lost, ghost and false rules, it is confirmed that our framework can protect the association rules efficiently in the big data environment.

With the coming of the World Wide Web and the rise of web-based business applications and informal organizations, associations over the web create a lot of information on a daily basis. It is becoming more complex and critical task to retrieve exact information from web expected by its users. In the recent times, the Web has extended its noteworthiness to the point of transforming into the point of convergence of our propelled lives. The search engine as an apparatus to explore the web must get the coveted outcomes for any given query. The greater part of the search engines can't totally fulfill user’s necessities and the outcomes are regularly inaccurate and irrelevant. knowledge of ontology and history is not much personalization in the existing techniques. To conquer these issues, data mining systems must be connected to the web and one advanced powerful concept is web-page recommendation which is becoming more powerful now a day. In this paper, the design of a fuzzy logic classifier algorithm is defined as a search problem in the solution space where every node represents a rule set, membership function, and the particular framework behaviour. Therefore, the hybrid optimization algorithm is applied to search for an optimal location of this solution space which hopefully represents the near optimal rule set and membership function. In this article, we reviewed various techniques proposed by different researchers for web page personalization and proposed a novel approach for finding optimal solutions to search the relevant information..

: Due to the generic feature of AOMDV, it is widely used in Long-Term Evolution (LTE) networks such as the Internet which requires new challenges to deploy conversational real-time applications like VoIP (Voice over IP) and video conferencing. In AOMDV, the multiple routes between any source and destination pair are selected based on minimal hop count which does not ensure reliable video content delivery as the communicating nodes along the paths are prone to link failures and route breaks. To overcome such problems, the Link Reliable On-demand Multipath Distance Vector (LROMDV) is proposed by modifying the Ad hoc On-demand Multipath Distance Vector (AOMDV). The LROMDV uses an enhanced Cumulative Expected Transmission Count (enh-CETX) for selecting multiple routes between any source and destination pair to avoid link failures and route breaks in Long-Term Evolution (LTE) networks. The performance of LROMDV on H.264/MPEG-4 AVC video streaming under both Distributed Coordination Function (DCF) and Enhanced Distributed Coordination Function (EDCF) using NS2.34 and Enhanced EvalVid framework is evaluated and the simulation result shows the effectiveness of LROMDV.